UT theoretical chemistry code forum

Hi,

If the MEP has been converged using the NEB method, is it better for me to start the dimer method from the same POSCARs (the saddle point from the NEB method) in the subdirectories 01 and 02 or from the different POSCARs (the two images on the left and right of the saddle point along the MEP)?

Thanks.

If you have used the climbing image, you should not need to run a dimer calculation. But sometimes you might not want to wait for the entire band to converge, or you might want to reconverge your saddle point with different calculation parameters, such as the number of kpoints. In this case, it is a good idea to start a dimer from an neb run.

Try using the neb2dim.pl script. This will use a cubic interpolation to find the location of the saddle point, and generate two images (POSCAR files) on either side of that point along the minimum energy path (MEP). You can also do this, or something similar by hand using the interpolate.pl script. Make sure you don't put the same POSCAR in the 01 and 02 directories, or the dimer will not have any information about the lowest curvature direction along the MEP.